Re: /update/xml dropping exceptions

[Ryan McKinley <ry...@gmail.com>]

I posted an update to SOLR-85 last week...  I'm still feeling
embarrassed over this bug...  can someone either check SOLR-85 or add
the line:

        rsp.setException( e );

to line 84 of RequestBaseHandler.

thanks!
ryan



On 1/31/07, Ryan McKinley <ry...@gmail.com> wrote:
> the fix is in SOLR-85
>
> It happened because in one version we threw all exceptions to the top
> and deprecated rsp.setException() and then we decided to deprecate
> rsp.setException in a separate issue - apparently i didn't catch
> everything :)
>
>
>
> On 1/31/07, Yonik Seeley <yo...@apache.org> wrote:
> > I haven't looked into it yet, but it seems like any problems in a
> > request to /update/xml get lost somewhere... a positive response is
> > always returned.
> >
> > -Yonik
> >
>

Re: /update/xml dropping exceptions

[Chris Hostetter <ho...@fucit.org>]

: sounds fine to me... this code was copied directly from the old
: request handler.  For a while in SOLR-104, we removed the

Hmmm... i'm not seeing that. looking at hte diffs for
StandardRequestHandler and DisMaxRequestHandler i don't see where they
were ever rethrowing SolrException

I'll make this change now.


-Hoss

Re: /update/xml dropping exceptions

[Ryan McKinley <ry...@gmail.com>]

>
> shouldn't both situations be handled the same way?
>

sounds fine to me... this code was copied directly from the old
request handler.  For a while in SOLR-104, we removed the
setException()... when we put it back, it got lost.

Yes, it should catch all exceptions and then call:
  rsp.setException( e );

thanks
ryan

Re: /update/xml dropping exceptions

[Chris Hostetter <ho...@fucit.org>]

: I posted an update to SOLR-85 last week...  I'm still feeling
: embarrassed over this bug...  can someone either check SOLR-85 or add
: the line:

i haven't had a chance to check out all of SOLR-85, but just so i'm clear
on this particular mini-bug, why add rsp.setException( e ) in the catch(
Exception e) block but not to the catch( SolrException se ) block
(replacing the "throw se") ?

shouldn't both situations be handled the same way?


-Hoss